One Expert, Every Site: Remote Visual Assistance for Specialized Maintenance

In an earlier post we argued that off-the-shelf video-call apps stop working once a call becomes part of a business process instead of a meeting. Here's one of the clearest examples we've built: a platform for remote visual assistance on highly specialized, hard-to-reach installations — the kind of place where getting a qualified specialist physically on-site is expensive, slow, or sometimes just not possible on short notice. Offshore energy platforms and extraction sites are the obvious examples, but the same shape shows up anywhere specialized maintenance work is scattered across remote locations.

The problem: an expert can't be everywhere

Some maintenance work genuinely needs deep, specialized expertise — but the people qualified to do it are few, and the sites that need them are scattered and often remote. The traditional answer is either a specialist stationed at every site, which is expensive to justify for problems that come up occasionally, or one who travels between them, which is slow and leaves every hour in transit as an hour not spent solving anything.

The alternative is to keep a junior or generalist team on-site permanently, and bring the actual expert in remotely, on call, whenever a problem needs their judgment. Sometimes the fix doesn't even need the expert to do anything themselves; just to see clearly enough to talk a less experienced worker through it.

Why a generic video call isn't the best choice

The client didn't stop using Teams, Zoom, or Meet — those still handle every ordinary meeting they have. This platform exists for one specific, narrower situation: an active incident at a remote site, where the expert needs far more than a webcam pointed at a problem.

Ordinary meeting apps are built around a simple assumption: every participant controls their own media. You mute yourself, you pick your own camera, you decide what you share. That assumption is exactly backwards for this use case. During an incident, the on-site team doesn't need control over the call — they need the expert to see and hear whatever the expert needs, on the expert's terms, without anyone on-site having to manage call settings while they're also holding a wrench. The tool this needs is asymmetric by design; the tools built for meetings aren't.

What we built

Each remote site has its own pre-configured session, already wired to that site's fixed IP cameras and, increasingly, its real-time IoT data. When an incident is reported — there's a problem at Site A — the expert opens Site A's session directly. The site's cameras and live sensor data load automatically; the expert then calls or notifies the on-site team to join — each of them just opens a link, no account to create.

From there, the session is entirely centered on the expert:

• The expert sees and hears everyone connected: every fixed camera, every team member's device, at once.
• On-site team members only hear the expert by default; they don't hear each other.
• The expert can selectively unmute one person's audio just for themselves, or broadcast it to the whole session.
• The expert can push any video source, a fixed IP camera or any team member's live feed, to one specific person or to everyone, so a technician can be shown exactly the angle they need to act on.
• On top of all that, the expert can reach into any live camera feed directly: crop, zoom, rotate, reframe; without asking anyone to touch their device.

None of this is about replacing a meeting. It's about giving one person, remotely, the same situational awareness, and the same authority to direct people and equipment, they'd have if they were standing in the room themselves. A single incident can genuinely need several people acting at once, too: one checking a valve, another reading a gauge, both coordinated live by the same expert.

The economics

The value isn't really about the video call; it's about what the video call replaces. Instead of a specialist permanently stationed at every remote site, or spending days in transit between them, one expert can support multiple locations and teams in the same shift, from wherever they actually are. Junior staff or generalists stay on-site full-time and handle the routine work; the specialist's time, the genuinely scarce resource, only gets spent on the moments that actually need it. For sites that are expensive or slow to reach, cutting out the travel isn't a nice-to-have; it's most of the savings.

Where this is headed

The version we built already ties each site's session to that site's own fixed cameras. The roadmap goes further: cameras addressable dynamically, by GPS or by sector, so the expert isn't limited to whatever was already wired into a site's session. If a worker is looking at a panel in a specific room, the expert could pull in whatever camera happens to be nearest, even one that wasn't part of the original setup.

This is exactly the kind of requirement that doesn't show up in an off-the-shelf video tool's feature list, because it isn't a video problem first; it's a coordination problem, with a genuinely asymmetric shape, that a video platform happens to solve once it's built around how the work actually happens. It's also a reminder of something this whole series keeps coming back to: a "video call" can mean a meeting, or it can mean a fully custom tool built around one very specific job — knowing which one you actually need is most of the work.